Skip to content

PERF: assert_frame_equal / assert_series_equal #55971

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 9 commits into from
Nov 17, 2023
Merged

PERF: assert_frame_equal / assert_series_equal #55971

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
7aa2edf
improve perf of index assertions
lukemanley Nov 15, 2023
896bb74
whatsnew
lukemanley Nov 15, 2023
dca4df6
Merge remote-tracking branch 'upstream/main' into perf-assert-functions
lukemanley Nov 16, 2023
3e5a379
faster _array_equivalent_object
lukemanley Nov 16, 2023
9b94837
add comment
lukemanley Nov 16, 2023
3c94198
remove xfail
lukemanley Nov 16, 2023
2358b63
Merge remote-tracking branch 'upstream/main' into perf-assert-functions
lukemanley Nov 16, 2023
004423f
Merge remote-tracking branch 'upstream/main' into perf-assert-functions
lukemanley Nov 17, 2023
1faabfd
skip mask if not needed
lukemanley Nov 17, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion doc/source/whatsnew/v2.2.0.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -305,7 +305,7 @@ Other Deprecations

Performance improvements
~~~~~~~~~~~~~~~~~~~~~~~~
- Performance improvement in :func:`.testing.assert_frame_equal` and :func:`.testing.assert_series_equal` for objects indexed by a :class:`MultiIndex` (:issue:`55949`)
- Performance improvement in :func:`.testing.assert_frame_equal` and :func:`.testing.assert_series_equal` (:issue:`55949`, :issue:`55971`)
- Performance improvement in :func:`concat` with ``axis=1`` and objects with unaligned indexes (:issue:`55084`)
- Performance improvement in :func:`merge_asof` when ``by`` is not ``None`` (:issue:`55580`, :issue:`55678`)
- Performance improvement in :func:`read_stata` for files with many variables (:issue:`55515`)
Expand Down
17 changes: 4 additions & 13 deletions pandas/_testing/asserters.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -43,7 +43,6 @@
Series,
TimedeltaIndex,
)
from pandas.core.algorithms import take_nd
from pandas.core.arrays import (
DatetimeArray,
ExtensionArray,
Expand Down Expand Up @@ -246,13 +245,6 @@ def _check_types(left, right, obj: str = "Index") -> None:

assert_attr_equal("dtype", left, right, obj=obj)

def _get_ilevel_values(index, level):
# accept level number only
unique = index.levels[level]
level_codes = index.codes[level]
filled = take_nd(unique._values, level_codes, fill_value=unique._na_value)
return unique._shallow_copy(filled, name=index.names[level])

# instance validation
_check_isinstance(left, right, Index)

Expand Down Expand Up @@ -299,9 +291,8 @@ def _get_ilevel_values(index, level):
)
assert_numpy_array_equal(left.codes[level], right.codes[level])
except AssertionError:
# cannot use get_level_values here because it can change dtype
llevel = _get_ilevel_values(left, level)
rlevel = _get_ilevel_values(right, level)
llevel = left.get_level_values(level)
rlevel = right.get_level_values(level)

assert_index_equal(
llevel,
Expand All @@ -328,7 +319,7 @@ def _get_ilevel_values(index, level):
diff = np.sum(mismatch.astype(int)) * 100.0 / len(left)
msg = f"{obj} values are different ({np.round(diff, 5)} %)"
raise_assert_detail(obj, msg, left, right)
else:
elif not left.equals(right):
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

do we know that .equals is the correct amount of strict?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think so, but after reviewing again I've reverted this line in favor of a lower level change to _array_equivalent_object.

This has the benefit of improving performance for more cases, e.g. not just the index values but in this case the series values:

import pandas as pd

N = 100_000

values = pd._testing.makeStringIndex(N).values

ser1 = pd.Series(values.copy())
ser2 = pd.Series(values.copy())

%timeit pd.testing.assert_series_equal(ser1, ser2)

1.92 s ± 56.1 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
22.3 ms ± 4.18 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)

# if we have "equiv", this becomes True
exact_bool = bool(exact)
_testing.assert_almost_equal(
Expand Down Expand Up @@ -592,7 +583,7 @@ def raise_assert_detail(
{message}"""

if isinstance(index_values, Index):
index_values = np.array(index_values)
index_values = np.asarray(index_values)

if isinstance(index_values, np.ndarray):
msg += f"\n[index]: {pprint_thing(index_values)}"
Expand Down